Homebrew oscilloscope in progress

[MrBungle]

Thanks to the generosity of another forum member I'm now in the possession of a working 3JP1 CRT so this project is on the go. I have a requirement for a simple triggered oscilloscope that works from a 12v source with around 5MHz bandwidth and a "reasonably calibrated" display so some finger in the air measurements can be done. This will be plugged into my likely never finished spectrum analyser and used as a general purpose LF scope. I've always been a proponent of building my own test gear for the sake of understanding but never got around to it until now, mainly due to time and budget constraints.

I'll post status on here and any construction photographs as I perform any work on it.

At the moment, this is going to be (and I apologise for the not very vintage approach) transistorised albeit a fully discrete design along the lines of a 1960's Tektronix 321A without the Nuvistors. Nice metal canned 2n2907 and 2n2222s which I have in abundance are the primary devices used for low voltage parts of the circuit. The general plan is:

1. Build a power supply to get the CRT anode and plate potentials. Currently this is going to be a self-wound transformer driven using a royer converter with a linear pre-regulator. Not wonderfully efficient but easy to manage. This will provide the acceleration and deflection potentials (+/-2kv and +/-350v) as well as the analogue voltages (+/-15v). I haven't wound a transformer of this scale or voltage before so it's an interesting project on its own. I expect a couple of failures in the process. I've ordered the cores and bobbin etc from CPC and the wire from scientific wire. I'll then set up the tube bias conditions, focus controls etc (I'm still working this one out if I'm honest)

2. Build horizontal and vertical push-pull deflection amplifiers. These are simple push-pull amplifiers with high voltage transistors at the moment. I haven't actually built or tested them yet.

3. Build vertical amplifier. This is a simple FET buffered amplifier with a phase splitter and follower with a switchable 10mV/100mV front end input. I'll worry about attenuators later which will be integrated into the switching scheme. I've already designed and tested this up to 20MHz. The bandwidth will be completely ruined by parasitics in the later stages however so I'm expecting this to fall to around 5MHz through all stages. Then it's down to playing around with peaking coils and other hacks to get some extra bandwidth out of the whole thing.

4. Build horizontal amplifier. This is designed and relatively well tested on a breadboard at least. It's a simple triggered multivibrator driving a current source ramp generator with a follower and phase splitter to drive the deflection amplifiers. The calculations I did were way off as it was done at 2AM so back to the drawing board for those but the circuit fundamentally works.

5. Stick it in a box.

Yesterday I got the CRT on my power supply and tested the heater. Relatively boring action shot included but it confirms the current required for the heater winding on the transformer is as per the datasheet for the tube and that the tube is good.



Fingers crossed time is on my side

[Mooly]

You might be interested in the Practical Wireless 'Purbeck' 'scope. This used a 3BP1 CRT and had a 6Mhz bandwidth. I built this at the time in the Watford Electronics case featured on the cover.

http://americanradiohistory.com/Arch...PW-1978-04.pdf

[KeithsTV]

There was a thread on the Purbeck a few years ago with all the PWs plus the followup.

https://www.vintage-radio.net/forum/...ad.php?t=65714

Keith

[Radio Wrangler]

London's out of my usual travel area, but there's a spare crt and matching mumetal shield & mountings in my attic if you want. 8x10cm flat face graticule, runs about 7 to 10kV on the PDA. dome-mesh electron lens so good sensitivity on Y plates and good to 100MHz.

David

[MrBungle]

Thanks for the offer. 7-10kv and 100MHZ is out of range for me at the moment though. I'll leave that sort of acceleration potentials and bandwidth to Tektronix et al

Thanks also for the PW purbeck articles links and the thread. It was actually the purbeck scope that inspired this originally many years ago when I found the April 1978 issue at a car boot sale in the early 1990s. Plus the rather beautiful insides of a Tektronix 321: http://www.helo.de/helo/tek/portable...321a_vor_e.htm

This is an independent design however designed from scratch with easy to obtain parts at the moment.

[Diabolical Artificer]

Yes those early Tek scopes are a work of art, I have a scrap Tek 545 if you want the ceramic wafer's and I have a few small scope tfmr's valve and transistor type if you can use them.

Miguel from Cuba built his own scope a year or so ago - https://www.vintage-radio.net/forum/...ad.php?t=81184

Andy.

[MrBungle]

I've actually got half a 545 and a 1A1 as well. It came in pieces, tubeless and missing bits. I'm recycling the ceramic strips for this.

Transformer parts other than one reel of wire arrived yesterday afternoon so I'm going to wind the primary windings this weekend, the heater winding and the analogue low voltage. Will do the two high voltage ones later when I've found my kapton.

Thanks for tbe link to Miguel's scope. This is a big project to bite off so I'm glad to see other people's

[Radio Wrangler]

Your first scope is always a big deal. I did mine when I was about fourteen, it used a 5BP1 tube and mains-driven EHT. Everyone on here says how dangerous mains EHT is, but I didn't know at the time. I just treated it with appropriate care. I'd bought a job-lot of EF91s so they got used for everything.

David

[David G4EBT]

Regarding the PW Purbeck, the 6-part series ran from April to September 1978 PWs and all of the issues can be found here:

http://www.americanradiohistory.com/...s_Magazine.htm

I built the ‘scope and it worked well - I still have it, but haven’t used it for some time as I have a Hameg 203-7, so the Purbeck is redundant. I started building PW projects as a schoolboy in the early 1950s, and it wasn’t long before bitter experience taught me that when projects I’d built faithfully following the instructions to the letter failed to work, that invariably it wasn’t due to ineptitude on my part, but to errors on PW’s part. Not for nothing did PW become known as ‘Camm’s Comic’.

Hence, I waited some months before embarking on building the Purbeck, steadily getting the bits together, knowing full well that with a project of that complexity, it was almost inevitable that errors would creep in, as they did. Lo and behold, in January 1979 there was a 3-page follow-up of errors and omissions. I made a number of economies, without which I wouldn’t have been able to have afforded to build it. Namely, I etched all my own PCBs, and made my own cabinet from sheet steel bought from a motor spares shop. Back then, is pre CAD days, I couldn’t produce the acetate front panel so bought it from PW.

By the standards of the day, the Purbeck was a cheaper option than a commercial scope, which - for all but the well off - would have been beyond the pocket of hobbyists, but that said, the Purbeck was- in real terms - an expensive project. I’ve still got the Purbeck price list from Watford Electronics. The total cost of parts including 8% VAT was £80.33. Even if the parts could be obtained today for that price, it would be a non-starter when you can get an excellent 20 - 30 MHz dual trace analogue ‘scope for less than that – typically £60 - £75 for something such as the Hameg 203-7, with a built-in component tester. But when you adjust the price of the kit, and the more expensive individual parts from Watford Electronics to take account of inflation, it puts into a proper perspective just what an expensive project it was.

To put the cost into perspective, here are some original prices and what they’d equate to today at 2016 prices, with 2016 equivalent in brackets:

Total kit price 1978: £80.43. (£451.45)
Case 1978: £15.50 (£87.00)
PCBs 1978: £14.87 (106.05)
Mains TX: £5.95 (33.44)
3BP1 CRT: £7.75 (43.56)
LM304 regulator £2.40 (13.49)
PW acetate front panel: £2.25 (12.65).

Inflation data taken from here: http://inflation.stephenmorley.org/

I also made the 10:1 probe which I no longer have, and when later there was a matching dual trace add-on unit by the same author (Ian Hickman), I etched the PCB, bought all the components, made up the ‘switchery’ and made a case. Before I could build it I acquired a fully working time-expired Gould OS300 scope put in a skip at work, so the Purbeck become redundant. I binned the dual trace case & PCB I think, but I still have most of the bits for that somewhere. (I also bought a Purbeck non-worker for more than I should have done, for spares if needed).

Happy days!

Have fun building the homebrew scope. I guess that those who embark on such a project these days will do so more from interest and enjoyment rather than necessity, so will most likely already have another scope to help with fault tracing on the homebrew one if need be!

[MrBungle]

Yes I've seen that before, once to my cost (angry fist shake in the direction of EPE). Most of the magazine projects were documented from one prototype grade build from experience. They just happened to mostly work once. I tend to look in the direction of service manuals for project inspiration as they have usually been revised several times before publication and thr kinks have been worked out of the designs. For example the PRC-77 technical manual, which I've always found fascinating, documents on several of the RF modules that component hand selection is required. The approach is usually try value X and if that doesn't work, swap out a resistor for value Y. Also the designs aren't amorphous blobs of components seemingly strung together with disregard for modular isolation.

This is definitely a leisure project. I have a perfectly good digital Tek and analogue Philips and Telequipment scopes already. However none of them run off 12v which would be handy because I'm in the process of planning a DC power system for the house as well as I'll be moving out into the sticks next year.

As for mains EHT, back in the distant past I managed to knock myself off a lab stool with a 2kv voltage source and ended up with a fortunately minor head injury so I have a healthy respect for EHT potentials. There is some motivation with this scope and the 12v requirement to isolate the EHT portions so it can be treated as a mostly safe low voltage instrument.

[MrBungle]

Minor update: managed to wind the primary and feedback windings and got the royer converter to do its thing. Managed to blow up a power transistor in the process as I wasn't paying attention to the job quite as much as i probably should have. Slightly behind schedule and not looking forward to the HV windings but I shall prevail. Very much liking these newfangled ferrite based transformers. No more bashing in varnished laminated ick!

[MrBungle]

Transformer winding complete followed by an instant failure. I wound the analogue rail windings and they are fine. I get about 1.5v higher than expected under load. The HV winding flashed over instantly with an unpleasant crack sound though and now no longer works. I am going to unwind and rewind it again tonight, probably separating alternate layers with kapton tape to keep higher potentials far away from each other.

For high voltages at a small current I find a lower winding voltage and a multiplier quite effective. Sometimes you can use an off the shelf transformer too.

[MrBungle]

Thanks for the tip. That is the plan in this case. The potential desired is a paltry +/- 2kv so not terrible. I was winding up to 1000-0-1000v then using a multiplier to double this to 2kv per rail but that gives a 2kv differential across the transformer which appears to have pushed the insulation a bit far. I'm going to try 500v and kapton each layer and use a quadrupler this time and go from there.

The only transformers I've wound before this were low voltage / RF toroids / rewind a mains secondary for another voltage.

[Ed_Dinning]

Hi Tape will be good, but try tapering the windings in (less turns on successive layers out) this will improve creepage and clearances and reduce risk of flashovers.

Ed

[MrBungle]

Good idea. Will try that. Thanks for the tip.

[MrBungle]

Minor update. Have rewound the whole transformer again so it runs at the correct voltage off an 8v supply which allows me to have some feedback headroom in the future and some voltage sag room when I put a pre-regulator on it, managed to break the core by accident (ferrite - grr!) and ordered another one that arrived Friday and blown up three power transistors. I've wound the heater winding, the HT winding, the deflection windings, the input windings, the feedback winding and the two rail windings all successfully. Used 2.1v/winding as it all divides up neatlyish. Looks a mess so I'm going to tidy it up before I post any photos.

A quick test with some suitably sized resistors and adding some capacitive feedback as it was kicking out pretty nasty square waves and we're in business. So, 12v DC in and:

- 6.3v AC RMS out (6.05 measured at 400mA draw = good)
- 500v AC RMS out (409 at 5mA draw = a little low but in spec)
- +70v AC RMS out (65 at 10mA draw = good)
- -70v AC RMS out (-69 at 10mA draw = good although I think I wound a couple of extra turns there)
- +20v AC RMS out (18.5 at 100mA draw = good)
- -20v AC RMS out (-18.6 at 100mA draw = good)

I changed the original woefully inadequate and positively incendiary BD139's for a pair of MJE3055T's bolted to a chunk of alu from a dismantled Chinese switch mode (the only thing in it worth keeping). Oscillation frequency is around 1.5KHz which means there are some nasty harmonics likely to appear so I'm concentrating on the power supply next with heavy filtering and regulation

Efficiency is around 70% which isn't bad for a first cut. I'm going to try and work out where this is going and see if I can up it 10% or so. The transistors it is not because they're not even warm. Transformer core doesn't get particularly warm either after I plucked up the courage to touch it and risk a zap from my bad winding method.

so next steps:

1. I'm adding a feedback-driven pre-regulator that uses the analogue +20 as a reference, divides it and throws it at an LM358 as an error amplifier, which in turn regulates the input to the royer converter with a TIP122 I found lying around.
2. Output filtering, regulation etc on the analogue and deflection lines.
3. Tidy it up.

Side issue. Managed to blow a small electrolytic capacitor up, merely a 1uF 63v. Bounced off my cheek when it went. Please wear safety goggles when doing this. I had a near miss.

Edit: I sold my TTi power supply last week as well so some budget has now appeared. The Heathkit one I have didn't knock out enough current to get this to run with loads attached so I grabbed a voltcraft 13.8v radio supply (FSP-1136) which can kick out 6A for just over £40. Turns out this is actually sufficient for most things I do. Also seems to make mincemeat of capacitors I install backwards!